Helsinki, Finland
Systems programmer, kernel developer. Having fun (no) developing xcore.
~littleendian assigned ~littleendian to #5 on ~littleendian/xcore-todo
Ticket created by ~littleendian on ~littleendian/xcore-todo
the interrupts are broken. they do not work as expected. something needs to be done, really.
Ticket created by ~littleendian on ~littleendian/xcore-todo
There is a constant flow of messages in xcore logging everything. We need to include logging scopes to xcore's internal modules. It doesn't matter if the logging implementation is good or not, as at the end of the day after initial bootstrap the logs will be managed by the syslogd anyway.
Comment by ~littleendian on ~littleendian/xcore-todo
Since commit 91e04e2c555736f4b35c797d371f4cd37ea4a4de there are a few things safety-checked:
- x86.cpu.__readVendor:
- Now checks if the provided
buffer
can contain 12 bytes or more.- x86.gdt.load:
- Checks whether or not the last element in the provided
table
is a null entry.- x86.seg.reload:
- Checks whether or not the actual segment selector is an executable one, if yes, fail.
NOTE: These safety-checks are ONLY AVAILABLE IN
Debug
andReleaseSafe
MODES. As a developer, if you build xcore in ReleaseFast or ReleaseSmall, your warranty is void, as the safety-checks will be omitted for performance reasons. If you develop like that, expect uncaught bugs.
Comment by ~littleendian on ~littleendian/xcore-todo
resolved since e33fa7ec.
REPORTED
RESOLVED FIXEDComment by ~littleendian on ~littleendian/xcore-todo
resolved since cf85f965
REPORTED
RESOLVED FIXEDimprovement added by ~littleendian on ~littleendian/xcore-todo
feature added by ~littleendian on ~littleendian/xcore-todo
~littleendian assigned ~littleendian to #3 on ~littleendian/xcore-todo